Repository Public Research preview
1{2 "schemaVersion": "1.0",3 "name": "@spboucher/openai-api-toolkit",4 "version": "1.0.2",5 "type": "skill",6 "description": "Write, review, debug and migrate OpenAI API code with model tables and a request linter built from live probes of 29 models. Covers Responses, Chat Completions, tools, costs and tested scaffolds.",7 "license": "MIT",8 "entrypoint": "SKILL.md",9 "keywords": [10 "openai",11 "gpt",12 "responses-api",13 "agents",14 "linter",15 "migration",16 "cost",17 "function-calling"18 ],19 "compatibility": [20 "claude-code",21 "codex",22 "cursor",23 "generic-agent-skills"24 ],25 "permissions": [26 "filesystem:read",27 "filesystem:write"28 ],29 "examples": [30 {31 "title": "Find the right endpoint and price a workload offline (Claude Code)",32 "client": "claude-code",33 "image": "examples/claude-code.png",34 "alt": "Terminal screenshot of a Claude Code session using openai-api-toolkit: the prompt, the tool or script calls, and the agent's answer.",35 "prompt": "Use the openai-api-toolkit skill (its scripts, no API key needed): lint a Chat Completions request with model gpt-5.3-codex, max_tokens 500 and temperature 0.5, say which endpoint this model needs, then estimate the cost of 1M input tokens and 200K output tokens on gpt-5.4, standard versus batch. Answer in at most 10 lines.",36 "caption": "The skill's scripts report that the model is served on the Responses API only, then price a workload at standard and batch rates. Real session, Claude Code 2.1.322 with Fable 5.2, with version 1.0.1 installed from this registry; inputs are test fixtures.",37 "capturedAt": "2026-11-10"38 },39 {40 "title": "Find the right endpoint and price a workload offline (Codex)",41 "client": "codex",42 "image": "examples/codex.png",43 "alt": "Terminal screenshot of a Codex session using openai-api-toolkit: the prompt, the tool or script calls, and the agent's answer.",44 "prompt": "Use the openai-api-toolkit skill (its scripts, no API key needed): lint a Chat Completions request with model gpt-5.3-codex, max_tokens 500 and temperature 0.5, say which endpoint this model needs, then estimate the cost of 1M input tokens and 200K output tokens on gpt-5.4, standard versus batch. Answer in at most 10 lines.",45 "caption": "The skill's scripts report that the model is served on the Responses API only, then price a workload at standard and batch rates. Real session, Codex CLI 0.167.1 with GPT-6.1-Sol, with version 1.0.1 installed from this registry; inputs are test fixtures.",46 "capturedAt": "2026-11-10"47 }48 ]49}50